Deep Reinforcement Learning via Past-Success Directed Exploration.
Xiaoming LiuZhixiong XuLei CaoXiliang ChenKai KangPublished in: AAAI (2019)
Keyphrases
- reinforcement learning
- active exploration
- exploration strategy
- action selection
- exploration exploitation
- exploration exploitation tradeoff
- autonomous learning
- function approximation
- model based reinforcement learning
- active learning
- model free
- balancing exploration and exploitation
- temporal difference
- state space
- learning algorithm
- optimal policy
- markov decision processes
- reinforcement learning algorithms
- success factors
- multi agent
- temporal difference learning
- learning process
- robotic control
- real world
- machine learning
- information systems
- website
- interactive exploration
- robot control
- database
- optimal control